Abstract Aggressive forms of breast cancer, often resistant to standard chemotherapy, are increasingly treated with targeted signaling pathway inhibitors. HER2 inhibitors are standard of care for treatment of HER2+ breast cancer, and inhibitors of PI3K as well as AKT both demonstrate promise for the treatment of triple negative breast cancer. While these therapies have shown improvements over traditional therapies, there are several mechanisms that can limit response, with accumulating evidence suggesting that a dominant mechanism limiting efficacy is the release of an intrinsic negative feedback loop that causes upregulation of the receptor tyrosine kinase human epidermal growth factor 3 (HER3). Excess HER3 forms heterodimers with HER2 monomers (in HER2+ breast cancer) or EGFR monomers (in triple negative breast cancer) and allows for continued growth pathway signaling. This dynamic upregulation of HER3 occurs within days of therapy initiation, and is not the result of acquired genetic mutations, but rather an intrinsic cellular response to attempt to maintain homeostasis. Furthermore, this pathway is variably active across patients, without any current method of predicting its activity prior to therapy initiation. Understanding if and when this resistance mechanism is active in a given patient started on targeted inhibitors remains clinically impractical, as it requires invasive tissue biopsy both before therapy initiation as well as during treatment. Such paired biopsies have associated patient-risk, and moreover a single-site biopsy does not reflect intrinsic tumoral heterogeneity. To rapidly and noninvasively identify breast cancer patients that will develop resistance to targeted inhibitors through increased HER3 expression and rationally guide subsequent therapeutic choices (such as HER3 inhibitors) to overcome this resistance, we propose to develop and utilize a clinically-translatable HER3 PET imaging paradigm. To this end we have developed a first-in-class HER3 targeted peptide for PET imaging. In the envisioned imaging paradigm patients would be imaged with HER3 PET prior to and again shortly after starting therapy, to assess for change in tumoral HER3 expression. We will investigate this approach in both established cell lines of HER2+ and triple-negative breast cancers. We will determine threshold levels of HER3 expression change, as assessed by HER3 PET SUV, that are predictive of a tumor overcoming targeted inhibition through increase in HER3 expression. We will demonstrate the ability of such imaging to predict resistance and subsequently guide adaptive therapy by testing in patient-derived xenografts, which more closely resemble the heterogeneity of clinical practice. This approach represents an evolution in the use of imaging to guide therapy on a personalized basis, providing immediate insight to mechanisms of therapy resistance and guiding alternative treatment strategies in a non-invasive manner, a major advancement over both standard anatomical imaging or invasive biopsy.